I’m bullish on AI. Anyone who’s been following our progress knows that to be true. I firmly believe that what we’ve seen so far is only just the beginning: that things are going to evolve from crazy to crazier, and sooner than we think. Call it AGI or not – it’s coming.
For consumers, this is exciting. The promise of AGI for everyday people is the presence of an actual Iron Man’s “Jarvis” in their pocket. When it comes to personal assistants, a genius capable of helping you with anything you can think of at a really high level is essentially what you want.
For businesses, however, it’s a much different story.
Put simply, the enterprise has far different needs than consumers. Enterprises are their own organisms, composed of thousands of intelligent individuals, all trying to work together.
To collaborate on complex work effectively, humans need policies, procedures, goals and guardrails. We need direction and hierarchy—governance and org charts—reinforced by deterministic processes and rules.
Without this guidance and clarity, we work inefficiently in the best case and destructively in the worst. Try hiring dozens of geniuses (humans) onto an existing team and assign them work, all simultaneously and without any structure – and what you’ll get is chaos.
It was for this reason—to manage intelligence, strategically utilize specialized talent in large groups—that mankind invented deterministic processes and policies.
You see determinism present in every effective large group of people throughout history, be they armies that win in battle—the Roman legions, the US Military—or organizations capable of shipping innovative new products or completing large-scale construction projects: from pyramids to iPhones. Each uses deterministic rules and guidelines and goals to translate talent and energy into efficiency and effectiveness.
AGI—a non-deterministic genius built in our image—requires just the same in order to work as part of a group.
Just as is the case with every intelligent human worker and every powerful tool that’s ever existed or that humans have ever created, to make AI work in the enterprise—let alone to ensure you use it safely—you must harness it with deterministic properties. You must be able to control what it does, how it does it, and, importantly, what it is not allowed to do.
Enterprise organizations will only embrace AGI with the right structures in place.
In the enterprise, both AI and AGI must bind a certain amount of non-deterministic capabilities—such as the ability to work autonomously and collaboratively with other AI tools over time to achieve complex, long-term initiatives—inside human-established deterministic rules.
This is a matter of strategy—it’s the only way to manage intelligence, as human history will attest—as well as safety.
Until recently, managing intelligence was synonymous with managing people, the most intelligent beings on earth. When that changes—when AI becomes smarter than humans—the need to deterministically manage intelligence will only increase, because both the constructive and destructive potential of that intelligence will be greater; after all, it’s never been the individual capacities of people nor the firepower of our tools that solely determines the effectiveness of any large group or organization, but how organizations leverage these assets to achieve their goals.
Enterprises who use AI so successfully won’t be those with the “smartest” AI, but those who use AI the smartest.